Microsoft has unveiled significant upgrades to its generative AI, Copilot, which now features voice, visual recognition, and emotional intelligence, transforming it into a more interactive companion for users.
Mustafa Suleyman, CEO of Microsoft AI, stated that these advancements allow computers to understand human language better, fundamentally changing the interaction paradigm between people and machines. 'These machines now understand our voice and can transcribe perfectly what we say,' he explained.
Copilot Voice, which allows users to engage in conversations, is being launched in Australia, Canada, New Zealand, the UK, and the US on iOS, Android, and the web. Initially available in English, it has demonstrated the ability to converse in other languages, including French and Spanish.
However, users are cautioned that while Copilot can provide well-formed responses, it may still generate false information, known as 'hallucinations.' Suleyman emphasized that the AI will not expand into sensitive areas, such as medical advice or legal warnings, until accuracy improves.
Another feature, Copilot Vision, enables the AI to analyze web content, including text and images, offering recommendations based on user preferences. This function is currently in beta and accessible through Copilot Labs in the US.
Additionally, Microsoft introduced Think Deeper, a feature designed for more complex reasoning tasks, and Copilot Daily, an audio summary of news and weather, which will eventually include emails and calendar events.